revert(rag): back out #2065 claim-leading condition binding after live canary regression

[BigSimmo]

Summary

• Reverts PR fix(rag): bind claim-leading for/in conditions in the high-risk trigger check (S1c follow-up) #2065 (fix(rag): bind claim-leading for/in conditions in the high-risk trigger check, commit 18a535cde, merge 964869fc3) — a single-commit revert per docs/rag-behaviour/ protocol after the post-merge live canary regressed.
• Evidence: eval-canary run 32097916649 (head 9904fbda8) failed the answer gate on agitation-im-po-route-short-terms ("What IM or PO options are listed for agitation?"): extractive route, identical topFiles, but routing reason became high_confidence_extractive_retrieval; claim_support_high_risk_gap; source_backed_review_fallback; extractive_quality_gate:template_like_answer; numeric_faithfulness_gate_source_gap → grounded:false, 0 citations, citation_failure_rate 0.0227. The prior canary (32052479537 on 084f63799) had the same case grounded with 5 citations. Retrieval golden set: 36/36, zero per-case rr regressions in both runs.
• Bisect (live probe, extractive path, deterministic — no model involved):aade20aba (parent of fix(rag): bind claim-leading for/in conditions in the high-risk trigger check (S1c follow-up) #2065) → 5 citations; 964869fc3 (fix(rag): bind claim-leading for/in conditions in the high-risk trigger check (S1c follow-up) #2065) → 0 citations; fix: clear the theme-transition timer leak and let issues:done close ULID-id ledger rows #2063, feat(rag): tag document-summary rows with a document_context provenance origin #2053 (G1), fix(rag): recover fast-route final-gate source gaps extractively (packet S1d) #2054 (S1d) heads all inherit the failure. The regressing change is the new condition-binding regex /^\s*(?:for|in)\s+([^,;.!?]{1,60}?)\s*,/gi in src/lib/rag/rag-claim-support.ts, which makes a claim-leading "For X, …" / "In X, …" bind X as a required condition token; on this extractive answer that flips a supported claim to high-risk-unsupported and cascades to an ungrounded review fallback.
• The immutable review record that fix(rag): bind claim-leading for/in conditions in the high-risk trigger check (S1c follow-up) #2065 added is deliberately kept (records are never deleted); only the source and its tests are reverted.

RAG impact: behaviour change — canary pair 32052479537 (last green, 084f63799) -> post-merge confirmation dispatch; this restores the claim-support behaviour that run 32052479537 measured green.

Verification

• Live probe on this branch: "routing_reason": "high_confidence_extractive_retrieval", "citation_count": 5 (was 0 on main).
• npx vitest run tests/rag-claim-support.test.ts — Tests 157 passed (157).
• npm run eval:rag:offline — Test Files 25 passed (25) · Tests 614 passed (614); Offline RAG fixture and production-contract checks passed.
• npm run check:rag:fixtures — Offline RAG fixture and manifest validation passed (36 golden cases, 25 suites).
• npm run check:production-readiness — READY: no blocking production-readiness failures.
• Prettier check on the two reverted files: All matched files use Prettier code style!

Risk and rollout

• Risk: low — returns rag-claim-support.ts to the exact state the last green canary measured. The R2/R3 work from fix(rag): verify imperative dosing claims against descriptive guideline norms (packet S1c, R2+R3) #2052 and fix: clear the theme-transition timer leak and let issues:done close ULID-id ledger rows #2063 remains in place.
• Rollback: re-apply fix(rag): bind claim-leading for/in conditions in the high-risk trigger check (S1c follow-up) #2065 together with a fixture reproducing this case, then re-canary.
• Provider or production effects: post-merge confirmation canary (owner-approved dispatch).

Clinical Governance Preflight

Complete this section when the change touches ingestion, answer generation, search/ranking, source rendering, document access, privacy, production env, or clinical output.

• Source-backed claims still require linked source verification before clinical use
• No patient-identifiable document workflow was introduced or expanded without explicit governance approval
• Supabase target remains Clinical KB Database (sjrfecxgysukkwxsowpy)
• Service-role keys and private document access remain server-only
• Demo/synthetic content remains clearly separated from real clinical sources
• Source metadata, review status, and outdated/unknown-source behavior remain conservative
• Deployment classification/TGA SaMD impact was checked when clinical decision-support behavior changed

Notes

• Follow-up for the S1c lane: the condition-first idea needs an offline fixture built from this exact extractive answer before it is re-attempted (queue via /issues).
• Ledger record for head afbf14737 under docs/branch-review-records/.

---

Opened by the RAG programme coordinator chat (Claude Code).

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.

[supabase]

This pull request has been ignored for the connected project sjrfecxgysukkwxsowpy because there are no changes detected in supabase directory. You can change this behaviour in Project Integrations Settings ↗︎.

---

Preview Branches by Supabase.
Learn more about Supabase Branching ↗︎.

[coderabbitai]

Warning

Review limit reached

You’ve reached a temporary PR review limit under our Fair Usage Limits Policy.

Your recent review volume is higher than typical usage, so adaptive limits are currently applied.

Next review available in:20 minutes

Limit details: You’ve used all 1 included review currently available under your plan. You completed 99 included PR reviews in the past 7 days; at that activity level, included reviews refill at 1 review per hour.

Your organization has reached its usage spending cap. Adjust your spending cap in the billing tab.

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: ca6c0306-9465-4655-89d9-e353c87af6e2

📥 Commits

Reviewing files that changed from the base of the PR and between 981d85d and fa87b77.

📒 Files selected for processing (3)

• docs/branch-review-records/a843a22bab4d2c65ae47bc683f553fb5e5726f2fe111fb3f93fa9386e7b46b20.record.md
• src/lib/rag/rag-claim-support.ts
• tests/rag-claim-support.test.ts

---

_{Comment @coderabbitai help to get the list of available commands.}